Vol. 15, No. 1 Printed in U.S.A.

INFECTION AND IMMUNITY, Jan. 1977, p. 132-137 Copyright X) 1977 American Society for Microbiology

Induction of Escherichia coli and Vibrio cholerae Enterotoxins by an Inhibitor of Protein Synthesis MARK LEVNER,* FRANK P. WIENER, AND BENJAMIN A. RUBIN

Department of Research and Development, Wyeth Laboratories, Philadelphia, Pennsylvania 19101

Received for publication 6 July 1976

Enterotoxigenic Escherichia coli or Vibrio cholerae 569B (Inaba) grown in the of the antibiotic lincomycin, an inhibitor of protein synthesis, produced elevated levels of heat-labile enterotoxin or choleragen, respectively, as assayed by both vascular permeability factor and capacity to elicit fluid accumulation in rabbit ileal loops. This induction of enterotoxin did not reflect either a coupling of lincomycin resistance with increased enterotoxigenicity or an effect of lincomycin on cellular release of enterotoxin, since spontaneously isolated lincomycin-resistant mutants ofboth E. coli and V. cholerae still required lincomycin for induction, and large increases in E. coli permeability factor activity were found intracellularly as well as extracellularly. After the period of exponential growth, E. coli became refractory to induction by lincomycin, although most of the induced enterotoxin activity appeared only after this period. No increase in copy number of the enterotoxin plasmid in E. coli 711 (P307) was found in induced cells by analysis of deoxyribonucleic acid reassociation kinetics. These and other data suggest that synthesis of enterotoxin, or at least its accumulation, is normally limited by cellular factors whose synthesis is preferentially inhibited by lincomycin. A possible connection between this phenomenon and lincomycinassociated diarrhea is considered. presence

The antibiotic lincomycin is a potent inhibitor of protein synthesis in gram-positive bacteria at concentrations on the order of 1 ,tg/ml. It also inhibits protein synthesis in Escherichia coli, but only at concentrations about 100-fold higher. The basis of this difference in sensitivity is unknown, but it appears to be reflected in cell-free protein synthesis as well as in intact cells (3). In vitro, lincomycin blocks the peptidyltransferase function of 50S ribosomal subunits, but the mechanism of inhibition in whole cells is not yet established (11). We find that lincomycin acts as an inducer of enterotoxin activity in both enterotoxigenic E. coli and Vibrio cholerae. This phenomenon may provide a useful probe for analyzing the regulation of synthesis, about which little is presently known, of these enterotoxins, and it suggests that the E. coli and V. cholerae enterotoxins, which cross-react immunologically and are similar in their modes of action (13), may also be similar with regard to the mechanisms which regulate their production. This finding may also have relevance for the understanding of lincomycin-associated diarrhea, a severe and frequent response to oral lincomycin therapy

port, and B. A. Rubin at the 1976 Annual Meeting of the American Society for Microbiology, 2-7 May, Atlantic City, N.J.) MATERIALS AND METHODS

Bacterial strains. Enterotoxigenic E. coli strains used included H197 and H10407, isolated from human subjects with cholera-like diarrhea and provided by R. Rappaport, and 711 (P307) and 711 (P155), K-12 strains carrying enterotoxin-specifying (ent) plasmids derived from E. coli strains of porcine origin and obtained from S. Falkow, who also provided the isogenic non-enterotoxigenic ent- strain 711. Vibrio cholerae 569B (Inaba) was also provided by R. Rappaport. Lincomycin-resistant (Linr) mutants are spontaneous isolates resistant to at least 300 ug (E. coli) or 60 ug (V. cholerae) of lincomycin per ml under our growth conditions, obtained by serial transfer of liquid cultures at increasing concentrations of lincomycin, followed by purification of resistant clones. The Linr phenotype is stable in the absence of the drug. Media and growth conditions for enterotoxin production. Overnight cultures grown in 2% (wt/ vol) peptone and 0.5% (wt/vol) NaCl at 37°C without aeration were diluted 1:1,000 in 40 ml of yeast extract-supplemented Casamino Acids-salts medium (9). (synyeast) (6), at times containing lincomycin (Lin(A preliminary report of this and related cocin, Upjohn) at various concentrations, and were work was presented by M. Levner, R. Rappa- rotated vigorously in 500-ml Erlenmeyer flasks at 132

VOL. 15, 1977

ENTEROTOXIN INDUCTION BY LINCOMYCIN

3700. In our hands, heat-labile E. coli enterotoxin, assayed as described below, appears to be released into the growth medium primarily during cell lysis, since its release approximately parallels that of (galactosidase (M. Levner and R. Rappaport, unpublished data), and we routinely assayed extracellular enterotoxin activity after a 48-h incubation. This procedure was also used for V. cholerae. Enterotoxin assays. (i) Rabbit skin test. Portions of bacterial cultures were cleared by centrifugation at room temperature, and supernatants were filtered through 0.45-jum (E. coli) or 0.22-Am (V. cholerae) membrane filters (Millipore Corp.) prior to assay for vascular permeability factor (PF) by the rabbit skin test, performed essentially as described by Evans et al. (5). PF activity, which forE. coli is a measure of heat-labile enterotoxin under these assay conditions (5), is expressed as the concentration (ml-') of 4-mm bluing doses. Each figure is the average of data obtained with two rabbits. The ratio of lincomycin-induced PF activity to uninduced activity varied from 10- to 80-fold for E. coli H197 Lin! and from 4- to 330-fold for V. cholerae 569B Linr in at least five repetitions of these experiments, in each of which induced and noninduced cultures were grown and assayed in parallel. (ii) Beal loop assay. Ligated rabbit ileal loops were injected with 1.0-ml portions of enterotoxin preparations at appropriate dilutions, and accumulation of fluid was measured 18 h later as described by Rappaport et al. (12). Estimation of relative plasmid DNA concentrations by reassociation kinetics. E. coli 711 (P307), which carries an enterotoxin-specifying plasmid (ent) and no other plasmids (7, 14), was grown in synyeast medium containing [methyl-3H]thymidine (20 ,uCi/ml, 60 Ci/mmol; Schwarz/Mann). Labeled ent deoxyribonucleic acid (DNA) was purified by preparation of cleared lysates (4) of these cells during exponential growth, followed by fractionation of supercoiled DNA in density gradients containing 50% (wt/wt) CsCl and 200 ;tg of ethidium bromide per ml, which were spun in a Beckman 40 rotor at 1500, 38,000 rpm, for at least 48 h. Pooled plasmid fractions were extracted twice with isopropanol and dialyzed extensively at 4°0 against 0.42 M NaCl. Electron microscopic examination of ent DNA preparations by the method of Inman and Schn6s (8) revealed uniform populations of circular doublestranded DNA with the expected molecular weight (about 60 x 106). The specific activity of 3H-labeled ent DNA was taken to be the same as that of the chromosomal DNA of the same cultures, since labeling was continued for at least four generations. The latter was determined by extraction of DNA from pelleted material remaining after preparation of cleared lysates, essentially as described below, and assuming a specific extinction coefficient at 260 nm of 20 ml/cm per mg. The specific activity of ent DNA varied from 60,000 to 170,000 cpm/;jg. Unlabeled total cellular DNA was extracted from E. coli 711, 711 (P307), and 711 (P307) Lin' grown to either midlog or stationary phase under the conditions described above for enterotoxin production, with the Linr strain either in the presence or ab-

133

sence of lincomycin (250 jg/ml). The extraction procedure, based on that of Marmur (10), consisted essentially of treatment of lysozyme-ethylenediaminetetraacetic acid-disrupted cells with 1% sodium dodecyl sulfate at 60°C for 10 min, followed by the addition of NaClO4 to 1.0 M and extraction with chloroform-isoamyl alcohol (24:1) (vol/vol). After dialysis overnight at 40C against 0.42 M NaCl, digestion was carried out with 50 Ag of boiled pancreatic ribonuclease (Calbiochem) per ml at 3700 for 30 min, followed by further extraction and dialysis with chloroform-isoamyl alcohol and 0.42 M NaCl, respectively. DNA reassociation reactions were carried out at 750C in 0.42 M NaCl, using 3H-labeled ent DNA in trace amounts, such that its renaturation was negligible in the incubation periods used, in the absence of added DNA. Unlabeled total cellular DNA was added to concentrations at least 750 times that of3Hlabeled ent DNA. All D,.NAs were sonically treated for 5 min with a Branson Sonifier at 150 W in an ice bath before denaturation by boiling for 10 min. The extent of reassociation of homologous singlestranded DNA molecules was determined by sensitivity of the labeled probe DNA to single-strandspecific S1 nuclease. Portions of DNA mixtures after various periods of incubation at 750C were diluted at least 30-fold into ice-cold S1 buffer (60 mM sodium acetate, 50 mM NaCl, 1 mM ZnCl,, pH 4.5) containing 20 .tg of denatured, sonically treated calf thymus DNA per ml, and stored at -20°C. All samples were then assayed simultaneously. S1 nuclease (Miles Laboratories, Inc.) was added to a concentration of 1,000 to 2,000 units/ml, and samples were incubated for 10 min at 370C. An equal volume of ice-cold 10% trichloroacetic acid was then added; after-at least 20 min on ice the samples were collected on Whatman GF/C filters, which were then dried and counted in Liquiflor-toluene scintillation fluid. Under these assay conditions, no detectable digestion of sonically treated native 3H-labeled total cellular E. coli 711 (P307) DNA occurred. A fraction of all 3H-labeled ent DNA preparations, which varied from 8 to 13%, remained resistant to S1 nuclease after boiling. The high degree of specificity in the assay is attested to by the data in Fig. 3, where it is shown that no detectable reassociation occurred in a mixture of total cellular DNA from E. coli 711, which lacks the ent plasmid, with trace amounts of 3H-labeled ent DNA. Reassociation data were plotted in terms of 1/f.,, the reciprocal of the fraction of labeled ent DNA remaining single stranded, as a function of time of incubation at 750C (t), and f. was corrected for the fraction of ent DNA remaining S1 resistant at t = 0 (8 to 13%) and S1 sensitive at apparent completion of the reaction (about 50%). Since, as mentioned above, native DNA was essentially completely S1 resistant, the reason for this high value of f. at the end of reassociation is not clear. Since 1/f. = 1 + KCot for an ideal second-order reaction and a given unique DNA sequence (2), where K is a constant, the slope of the reassociation curves was taken to be proportional to the concentration of ent DNA (Co) in

134

INFECT. IMMUN.

LEVNER, WIENER, AND RUBIN

the unlabeled total cellular DNA, since the concentration of labeled ent DNA was negligible.

RESULTS Induction of increased E. coli enterotoxin production by lincomycin is shown in Table 1. In experiment la, growth of E. coli H197 in medium containing 30 ,ug of lincomycin per ml resulted in at least a sevenfold increase in extracellular PF activity. At 100 ,ug of lincomycin per ml growth of E. coli H197 was severely inhibited, but cultures eventually attained a high cell density, suggesting that selection for lincomycin-resistant cells might also select for cells with increased enterotoxigenicity. However, isolations of spontaneous Line mutants, which are stably resistant to at least 300 gg of lincomycin per ml, indicated (as shown for one such mutant in Table 1, experiment lb) that this was not the case. Without the drug, no significant increase in PF activity was found, whereas in 100 to 500 jig of lincomycin per ml, at least 25fold enhancement was observed. Of 11 such mutants independently isolated from strain H197, 10 showed at least a fivefold increase in PF production when grown with 250 ug of lincomycin per ml. Three other enterotoxigenic E. coli strains were also examined for lincomycin induction, and all showed increased PF activity when grown in the presence of 50 ,ug of lincomycin per ml (Table 1, experiment 2). V. cholera enterotoxin (choleragen) is also susceptible to lincomycin induction. A spontaTABLE 1. Lincomycin induction of enterotoxin (PF)

Expt

la lb

2

3

Bacteria E. coli H197 E. coli H197 E. coli H197 E. coli H197 Linr E. coli H197 Linr E. coli H197 Linr E. coli H197 Linr E. coli H10407 E. coli H10407 E. coli 711 (P307) E. coli 711 (P307) E. coli 711 (P155) E. coli 711 (P155)

Lincomycin PF activity 0 10 30 0 30 100 500 0 50 0 50 0 50

V. cholerae, Inaba 0 569B V. cholerae, Inaba 0 569B Lanr V. cholerae, Inaba 50 569B Linr a See Materials and Methods.

Induction of Escherichia coli and Vibrio cholerae enterotoxins by an inhibitor of protein synthesis.

Vol. 15, No. 1 Printed in U.S.A. INFECTION AND IMMUNITY, Jan. 1977, p. 132-137 Copyright X) 1977 American Society for Microbiology Induction of Esch...
982KB Sizes 0 Downloads 0 Views